Control over the Structural and Optical Features of Nanoparticle

Langmuir 2009, 25, 2443-2448

2443

Control over the Structural and Optical Features of Nanoparticle-Based One-Dimensional Photonic Crystals Mauricio E. Calvo, Olalla Sa´nchez-Sobrado, Silvia Colodrero, and Herna´n Mı´guez* Instituto de Ciencia de Materiales de SeVilla, Consejo Superior de InVestigaciones Cientı´ficas (CSIC), Ame´rico Vespucio 49, 41092 SeVilla, Spain ReceiVed September 12, 2008. ReVised Manuscript ReceiVed NoVember 5, 2008 Herein we present a detailed analysis of the effect of the spin-coating protocol over the optical properties of nanoparticle-based one-dimensional photonic crystals. Based on these results, we provide a reliable synthetic route to attain high-quality porous multilayers in which the effect of imperfections is minimized and whose Bragg diffraction can be precisely tuned over the entire visible and near-infrared spectrum. We present a systematic study of the effect of the acceleration ramp and final rotation speed over the structural and optical quality of these materials. This allows us to relate the structural variations observed with the different relative importance of fluid flow and solvent evaporation on the thinning of each layer in the stack for the different deposition conditions employed.

Introduction A one-dimensional photonic crystal (1DPC) is a structure that presents a periodic modulation of the refractive index in one direction of the space.1 The interference effects associated with these periodic dielectrics gives rise to the opening of a photonic band gap whose effect is detected as a maximum in the specular reflectance spectrum of the structure. The presence of this Bragg peak is evidence that the propagation of photons into the material is forbidden. If this peak arises in the visible region of the electromagnetic spectrum, the 1DPC presents bright colors. The easiest way to create these structures is to alternate layers of materials with different refractive index. To achieve high optical quality stacks, individual layers must be uniform and interfaces between neighboring ones should be smooth. Due to the large dielectric constant contrast between TiO2 and SiO2, this is one of the preferred pairs of base materials to build such structures. A wide variety of techniques such as pulsed laser deposition, reactive sputtering, or different types of chemical vapor deposition techniques have been successfully applied to obtain Bragg reflectors.2 Among the materials chemistry-based approaches, wet deposition methods such as dip coating or spin coating of flat substrates with SiO2 or TiO2 sol-gel precursors have also been employed with satisfactory results.3,4 Recently, a series of mesoporous 1DPC films have been obtained by a “bottom up” approach based on dip- or spincoating different types of suspensions5 or precursor solutions6,7 of TiO2 and SiO2. Preliminary studies on the environmental response of these materials already demonstrated their great potential as base materials for gas and liquid detection.8,9 In one of these approaches,5 nanoparticles of each one of the abovementioned metal oxides are alternately deposited onto a flat

substrate to attain a multilayer structure that displays bright structural color and that possesses porosity controllable through both the degree of aggregation and the particle size of the precursor suspensions. This concept has been further extended to build photoconducting Bragg mirrors made of alternate layers of nanocrystalline titania in which the refractive index is spatially modulated only through the degree of porosity of each layer.10 This constitutes the first example of a multifunctional 1D photonic crystal in which photonic, photoconducting, and diffusion characteristics are combined to yield a material in which the photoelectric response can be tuned through the photonic properties of the ensemble. Furthermore, nanoparticle-based 1DPC have been proven to provide significant enlargement of the solar-to-electric power conversion efficiency in dye-sensitized solar cells in which they are easily built up.11 In this case, they operate as porous and highly reflecting coherent scattering layers that enhance the absorption of well-defined frequency ranges within a coupled electrode with no deleterious effect on the photovoltage.12 These examples illustrate the potential of these versatile architectures in which porous interconnected mesostructure is combined with a tunable Bragg diffraction along the visible and near-infrared spectra. In all these previous works, nanoparticle 1DPCs displaying different colors were prepared by modifying the sol particle concentration and by using a fixed final rotation speed.5 However, the range of suitable concentrations to achieve optical quality materials is not wide since we found that high concentrations gave rise to particle aggregation in the precursor suspensions and low ones yielded poor layer adherence. Besides, changing particle concentration could easily modify the rheological properties of the suspension. This causes the relation between mass fraction and layer thickness to be

* To whom correspondence should be addressed. E-mail: hernan@ icmse.csic.es. (1) Joannopoulos, J. D.; Meade, R. D.; Winn, J. N. Photonic Crystals: Molding the Flow of Light; Princeton University Press: Princeton, NJ, 1995. (2) Macleod, H. A. Thin Film Optical Filters, 3rd ed.; Institute of Physics: London, 2001. (3) Rabaste, S.; Bellessa, J.; Brioude, A.; Bovier, C.; Plenet, J. C.; Brenier, R.; Marty, O.; Mugnier, J.; Dumas, J. Thin Solid Films 2002, 416, 242. (4) Almeida, R. M.; Rodrigues, A. S. J. Non-Cryst. Solids 2003, 326, 405. (5) Colodrero, S.; Ocan˜a, M.; Miguez, H. Langmuir 2008, 24 (9), 4430. (6) Choi, S. Y.; Mamak, M.; Freymann von, G.; Chopra, N.; Ozin, G. A. Nano Lett. 2006, 6, 2456. (7) Fuertes, M. C.; Lo´pez-Alcaraz, F. J.; Marchi, M. C.; Troiani, H. E.; Mı´guez, H.; Soler Illia, G. J. A. A. AdV. Funct. Mater. 2007, 17, 1247.

(8) Fuertes, M. C.; Colodrero, S.; Lozano, G.; Gonza´lez-Elipe, A. R.; Grosso, D.; Boissie`re, C.; Sa´nchez, C.; Soler-Illia, G. J. de A. A.; Mı´guez, H. J. Phys. Chem. C 2008, 112, 3157. (9) Colodrero, S.; Ocan˜a, M.; Gonzaı`lez-Elipe, A. R.; Mı´guez, H. Langmuir 2008, 24, 9135. (10) Calvo, M. E.; Colodrero, S.; Rojas, C. T.; Ocaña, M.; Anta, J. A.; Mı´guez, H. AdV. Funct. Mater. 2008, 18 (18), 2708. (11) Colodrero, S.; Mihi, A.; Ha¨ggman, L.; Ocaña, M.; Boschloo, G.; Hagfeldt, A.; Mı´guez, H. AdV. Mater. 2008, published online on Dec. 16, 2008, DOI: 10.1002/adma.200703115. (12) Colodrero, S.; Mihi, A.; Anta, J. A.; Ocan˜a, M.; Mı´guez H. J. Phys. Chem. C 2009, published online Jan. 5, 2009, DOI: 10.1021/jp8078309.

10.1021/la8030057 CCC: $40.75  2009 American Chemical Society Published on Web 01/20/2009

2444 Langmuir, Vol. 25, No. 4, 2009

complex13 and prevents accurate control over the thickness of the layers forming the 1DPC. Herein, we present a detailed analysis of the effect of the spin-coating protocol over the structural and optical properties of nanoparticle based one-dimensional photonic crystals. Based on these results, we provide a reliable synthetic route to attain high optical quality porous multilayers in which the effect of imperfections is minimized and whose Bragg diffraction can be precisely tuned over the entire visible and near-infrared spectrum. We present a systematic study of the effect of the acceleration ramp and final rotation speed over the structural and optical quality of these novel materials. This allows us to relate the structural variations observed with the relative importance of fluid flow and solvent evaporation on the thinning of each layer in the stack for the different deposition conditions employed. We find that control over the acceleration ramp employed is crucial for determining with precision the final properties of the multilayer. As far as we know, this work is the first comprehensive study of the influence of the deposition protocol on the properties of Bragg reflectors made by spin coating.

Experimental Section Particulate Suspensions. TiO2 nanoparticulated sols were synthesized using a procedure based on the hydrolysis of titanium tetraisopropoxide (Ti(OCH2CH2CH3)4, 97% Aldrich, abbreviated TTIP) as has been described before.14 Briefly, TTIP was added to Milli-Q water. The white precipitate was filtered and washed several times with distilled water. The resultant solid was peptized in an oven at 120 °C for 3 h with tetramethylammonium hydroxide (Fluka). Finally, the suspension obtained was centrifuged at 14.000 rpm for 10 min. SiO2 nanocolloids were purchased from Dupont (LUDOX TMA, Aldrich). Both suspensions were diluted in methanol to 4 and 2 wt % for TiO2 and SiO2 particles, respectively. Multilayer Films. Photonic crystals were built by an alternated deposition of TiO2 and SiO2 nanoparticulated suspensions, following a generic procedure previously reported by our group.5 These sols were deposited over zero fluorescence glass using a spin coater (Laurell WS-400E-6NPP) in which both the acceleration ramp and the final rotation speed could be precisely determined. The first layer was deposited using 250 µL of SiO2 sol and the substrate was tilted and rotated to allow the suspension to cover the total glass surface. Then the sample was accelerated up to a different final speed using also several ramps in each case, to test the effect of these parameters. Final speed ω was chosen between a nominal value of 2000 and 8000 rpm and accelerations (γ) were selected between nominal 1950 and 13650 rpm s-1. The total spin-coating process (ramping-up and final speed) is completed in 60 s. Afterward, the coated sample is maintained at 25 °C for 5 min in a closed chamber. Sequentially, another layer of a different type of nanoparticle is deposited following the procedure described above. The process is repeated until a total of six layers have been deposited. Optical Measurements. Reflectance spectra were performed using a Fourier transform infrared spectrophotometer (Bruker IFS-66 FTIR) attached to a microscope and operating in reflection mode with a 4× objective with 0.1 of numerical aperture (light cone angle (5.7°). Reflectance spectra were acquired from the geometrical center of the sample to the border of the glass substrate in a straight line to the center of the side with a periodicity of 2 mm. The images of the films were acquired using a digital camera adapted to a microscope. More thorough optical maps were performed when required.

Results and Discussion Microstructure and Optical Properties. The structure of the different sintered 1DPCs was studied by analyzing the cross (13) Nitta, S. V.; Jain, A.; Wayner, P. C.; Gill, W. N.; Plawsky, J. L. J. Appl. Phys. 1999, 86 (10), 5870. (14) Burnside, S. D.; Shklover, V.; Barbe´, C.; Comte, P.; Arendse, F.; Brooks, K.; Gra¨tzel, M. Chem. Mater. 1998, 10, 2419.

CalVo et al.

sections of multilayers deposited under different conditions. Field emission scanning electron microscopy (FESEM) images of cleaved films made of three SiO2-TiO2 bilayers are displayed in Figure 1. The different morphology of the SiO2 and TiO2 nanoparticles employed allows the two different kinds of layers in the periodic arrangement to be clearly distinguised. SiO2 Ludox particles are spherical and larger than the irregular TiO2 crystallites. As expected, increasing the final rotation speed diminishes the final thickness of the layers, as Figures 1a and 1b illustrate. In these examples, the same SiO2 and TiO2 (4 and 2 wt % precursor suspensions) were deposited and spun at 2500 and 6000 rpm, respectively. The acceleration used was 8450 rpm s-1 in both cases. As we have pointed out before, two types of interfaces result from the deposition method employed. The surface between neighboring layers is flat when SiO2 nanoparticles are deposited onto the more compact TiO2 crystallites and is more diffuse when the opposite occurs since TiO2 crystallites are then able to penetrate a few nanometers into the larger pores of the Ludox packing. This can be clearly observed in the detail amplified in Figure 1c. Uniformity and Defects. To check the dependence of the quality of the multilayer with the experimental deposition conditions, we prepared several tens of samples and performed an optical study by measuring and analyzing their reflectance using a microscope attached to a spectrophotometer. The main sign of the interference effects that occur as a consequence of the built up periodicity is a reflectance peak whose position and width depends on the refractive index and thickness of the layers as well as, on a second-order approximation, the number of unit cells. Secondary lobes result also from interference effects, but caused in this case by the finite size of the multilayer. These features can be compared between spectra attained from different spots of the same 1DPC to test its structural uniformity, since lack of layer thickness homogeneity along the substrate would result in variations of the optical response. Taking advantage of the radial symmetry of the spin-coating process, we collected reflectance spectra from 1 mm2 area spots every 2 mm from the geometrical center of the substrate to one of its sides in all samples. Repeatability of the average optical response was also confirmed in different multilayers prepared under similar conditions. In Figures 2a and 2b, we plot the results obtained for multilayers built at a fixed acceleration ramp of 3250 rpm s-1 and two final speeds, namely, 2000 rpm (Figure 2a) and 8000 rpm (Figure 2b). In addition, in Figure 2c and 2d we show the spectra obtained for samples made using an acceleration of 8450 rpm s-1 and final rotation speeds of 2000 rpm (Figure 2c) and 8000 rpm (Figure 2d). It can be straightforwardly concluded from the analysis of these results that best structural uniformity of the 1DPC is achieved when multilayer deposition is carried out using high final rotation speeds, and that it further improves if steep accelerations ramps are employed. In fact, we find a systematic lack of uniformity for 1DPCs deposited at final rotation speeds below 3000 rpm. The reflectance maximum intensity can be directly related to the number of imperfections in the sample since they give rise to diffusely scattered light that is actually removed from the reflected and transmitted beams. In our case, regardless of the uniformity of the response, the height of the peaks is never below 60% for 6 layer (3 unit cell) films, which is always close to the expected theoretical values.15 Among the defects typically caused by the process of spin coating of a particle suspension, the most relevant ones are those known as “striations” and “comets”. Striations are radially oriented undulations of the film and are (15) Calculations were performed using a scalar wave approximation, as described in refs 7 and 10.

Nanoparticle-Based 1D Photonic Crystals

Langmuir, Vol. 25, No. 4, 2009 2445

Figure 1. SEM images showing the cross section of periodic nanoparticle multilayers obtained at final speeds of (a) 2500 rpm and (b) 6000 rpm. Image displayed in (c) shows a detail of (a) in which the two different types of interfaces existing between SiO2 and TiO2 layers can be seen.

Figure 2. Specular reflectance taken from 1DPCs deposited using different spin-coating conditions: (a) ω ) 2000 rpm, γ ) 3250 rpm s-1; (b) ω ) 8000 rpm, γ ) 3250 rpm s-1; (c) ω ) 2000 rpm, γ ) 8450 rpm s-1; (d) ω ) 8000 rpm, γ ) 8450 rpm s-1. In each case, spectra taken at different distances are plotted using a different line color: black, 0 mm; red, 2 mm; blue, 4 mm; green, 6 mm; pink, 8 mm.

the result of the rotation of a surface cell pattern.16,17 Such pattern may be present in any liquid film with one of its surfaces exposed

to air as a result of the Marangoni effect.18 In such films surface tension gradients, which in our case could be originated by local

(16) Haas, D. E.; Birnie, D. P., III; Zecchino, M. J.; Figueroa, J. T. J. Mater. Sci. Lett. 2001, 20, 1763.

(17) Strawhecker, K. E.; Kumar, S. K.; Douglas, J. F.; Karim, A. Macromolecules 2001, 34 (14), 4669.

2446 Langmuir, Vol. 25, No. 4, 2009

CalVo et al.

Figure 3. Optical microscopy images of different multilayers built at different final and ramp accelerations. Each pair of images was obtained from multilayers deposited at the same final rotation speed and accelerations of γ ) 11050 rpm s-1 and γ ) 1950 rpm s-1 (top and bottom images in each case, respectively). (a) ω ) 8000 rpm, (b) ω ) 6000 rpm, and (c) ω ) 2000 rpm. In all cases the scale bar ) 1 mm.

fluctuations of the suspended particle concentration or the temperature, induce fluid flows that set up a cell surface pattern. Rotation of such pattern gives rise to the elongation of the cells, an effect that is more intense as the distance from the center increases, which yields the striation pattern typically observed in spin-coated films made of particles.19 Comets are abrupt disruptions of the film uniformity caused by the presence of large particle aggregates that may be already present in the precursor suspension or form during the spreading of the suspension on the substrate. These two types of imperfections can be readily seen when the multilayers are observed under the optical microscope. A series of pictures taken from 1DPCs deposited using different spin-coating protocols are displayed in Figure 3. Figures 3a and 3b show respectively a pair of images of 1DPCs prepared using ω ) 8000 rpm and ω ) 6000 rpm and accelerations of γ ) 11050 rpm s-1 (top images) and 1950 rpm s-1 (bottom images). In all cases, the occurrence and characteristics of the striations observed is barely dependent on the spin-coating protocol used. Contrarily, the density of comets increases dramatically when slow ramps are used. In our case, surprisingly, the amount of comets observed in films deposited simultaneously onto similar substrates and from the same precursor suspension changes with the conditions used in the spin-coating protocol, as we can see by comparing each pair of images shown in Figures 3a, 3b, and 3c. This indicates that such comets originate during fluid flow and solvent evaporation from the rotating film. Those phenomena could break the repulsive forces between particles, leading to formation of bigger aggregates. In the case of an impulsive acceleration (higher γ values) solvents abandon the suspension rapidly, which causes an abrupt increase of viscosity to decrease of fluid flow. Under these conditions, aggregation of the nanoparticles may be hindered and, consequently, the formation of comets. Optical microscopy also shows the large difference in color resulting from using different acceleration ramps to (18) Scriven, L. E.; Sternling, C. V. Nature 1960, 187, 186. (19) Rehg, T. J.; Higgins, B. G. AIChE J. 1992, 38 (4), 489.

Figure 4. Reflectance spectra obtained at final rotation speeds of (a) ω ) 2000 rpm and (b) ω ) 6000 rpm. Acceleration speeds, γ (rpm s-1) ) black, 1950; pink, 3250; teal, 4550; blue, 5850; red, 8450; green, 11050.

reach higher final rotation speeds (ω > 3000 rpm). Color remains invariable when the 1DPCs are deposited at low ω, as the images corresponding to a multilayer prepared at ω ) 2000 rpm and different γ shown in Figure 3c illustrate. Analysis of the Dependence of the Optical Response with the Spin-Coating Parameters. In Figures 4a and 4b we present two sets of reflectance spectra obtained from 1DPCs deposited at final rotation speeds ω ) 2000 rpm and ω ) 6000 rpm, respectively. For the sake of comparison, all spectra were taken from spots located at 2 mm from the geometrical center of the substrate on which the 1DPC is deposited. For each ω, the spectra measured for photonic crystal films deposited using different ramp accelerations comprised between γ ) 1950 rpm s-1 and

Nanoparticle-Based 1D Photonic Crystals

Langmuir, Vol. 25, No. 4, 2009 2447

Figure 6. (a) Bragg peak spectral position (nm) versus ramp stage acceleration γ. The different lines correspond to different final rotation speeds ω (rpm): black, 2000; red, 2500; green, 3000; blue, 3500; teal, 4000; pink, 6000; yellow, 8000. Figure 5. Reflectance spectra obtained at acceleration speeds (a) γ ) 1950 rpm s-1 and (b) γ ) 11050 rpm s-1. Final speeds, ω (rpm) ) blue, 2000; red, 2500; black, 3000; green, 3500; teal, 4000; pink, 6000; blue, 8000.

γ ) 11050 rpm s-1 are shown. These results illustrate in more detail what can be seen with the naked eye: The higher ω is, the stronger the dependence of the Bragg peak position on the acceleration ramp used. It shifts from longer to shorter wavelengths as the acceleration increases, and for ω > 6000 rpm, it can take any value in a 200 nm range within the visible spectrum. Contrarily, for the smaller final rotation speeds, the Bragg peak is barely dependent on the γ employed. On the other hand, in Figure 5 we show a set of reflectance spectra obtained using different final rotation speeds but fixed accelerations of γ ) 1950 rpm s-1 (Figure 5a) and γ ) 11050 rpm s-1 (Figure 5b). These graphs illustrate another clear trend observed in the course of the experiments: The higher γ is, the stronger the dependence of the Bragg peak position on the acceleration ramp used. At the lower acceleration rate, the Bragg peak maximum oscillates around 600 nm for all final speeds, whereas for the fastest ramps the reflectance peak shifts from 620 to 380 nm as final speed increases. These results serve as examples to highlight the extraordinary importance of the spin-coating protocol to determine the structural, and thus optical, properties of the 1DPCs. To facilitate the analysis of the results attained for all the conditions employed, the reflectance peak positions obtained for the 1DPCs are plotted as a function of the ramp stage acceleration for all final rotation speeds used in Figure 6. The length of the error bars are extracted from a statistical analysis of the position of the Bragg peak in various regions of each sample. Thus, these bars can be interpreted as an estimation of the uniformity of the optical response of the crystal. The set of graphs plotted in Figure 6 can be used as a guide to attain 1DPCs of high structural quality whose Bragg peak precisely matches the desired wavelength range. Hence, it can be seen that worse optical quality is attained when the slowest ramps (γ ) 1950 rpm s-1 and γ ) 3250 rpm s-1) are used, independently of the nominal final rotation speed. As the slope of the ramp increases, so too does the optical quality of the multilayer. In general, the optical quality attained is good when both γ g 4550 rpm s-1 and ω g 2500 rpm. A reason for this effect of ω and γ on the colloidal multilayer quality can be given in terms of the border effects usually observed in spin-coating processes, which is one of the main sources of

the lack of uniformity observed in spin-coated films. When the precursor suspension film is spread onto the substrate, the wet film arrives at the border of the substrate and bounces in the opposite direction, leading to thickening in the surrounding region. When ω and γ are sufficiently high, this effect is negligible and the border effect disappears. Regarding the control of the optical response, the results summarized in Figure 6 clearly show that when the final rotation speed is reached fast, be it due to the large γ (γ < 5850 rpm s-1) or the low nominal ω (ω < 3000 rpm) employed, then λBragg depends almost entirely on such ω. Contrarily, for slower ramps and larger final rotation speeds, which implies that it takes a longer time to reach ω, then γ largely influences λBragg. These effects can be understood in terms of the relative influence on the process of film thinning of viscous “fluid flow”, on one hand, and on the other hand, “solvent evaporation”, as explained in what follows. After the classical work of Meyerhofer,20 the thickness T of a film deposited by spin coating is expected to depend on ω thorough the expression T ∼ ω-b, where b is a constant known as the “spin parameter”, with a typical value of 0.5 for Newtonian fluids. Meyerhofer also introduced the concept of “crossover point”, which is the time at which the contribution of fluid flow becomes, due to the increasing viscosity of the forming film, so small that it equals that of evaporation. The relation T ∼ ω-0.5 is obtained considering that thinning is first and foremost caused by viscous fluid flow due to spinning and, to a minor extent, by solvent evaporation. Recently, Birnie et al. have theoretically demonstrated that this expression is valid only if the crossover point is reached when the fluid is already rotating at the final nominal speed ω.21 After their study, if most film thinning takes place during the ramp stage, i.e., the crossover point occurs before the nominal ω is reached, then the spin parameters should be b < 0.5. We can use this prediction to provide some insight into the spin-coating processes. In our case, in a first approximation, the position of the specular reflectance Bragg peak maximum, λBragg, depends linearly on the average refractive index and the lattice constant of the ordered multilayer. The former is determined by the porosity and refractive index of the layers, which in turn depend on the composition, shape, (20) Meyerhofer, D. J. Appl. Phys. 1978, 49, 3993. (21) Birnie, D. P.; Hau, S. K.; Kamber, D. S.; Kaz, D. M. J. Mater. Sci.-Mater. Electron. 2005, 16 (11-12), 715.

2448 Langmuir, Vol. 25, No. 4, 2009

Figure 7. Log-Log graph representation of the Bragg peak position against final rotation speed. Best fitting obtained in ramp, γ (rpm s-1): black, 1950; red, 3250; green, 4550; blue, 5850; teal, 8450; pink, 11050.

and packing of the nanoparticles that form them, hence being independent of ω and γ. The latter, however, is determined by the thickness of the two types of nanoparticle layers (SiO2 and TiO2), which should depend on ω through the expression T ) ω-b. Assuming that both kinds of layers depend on ω in a similar way, then we expect the Bragg peak maximum to follow the expression λBragg ∼ ω-b. So we can estimate the spin parameter of our multilayers by plotting λBragg versus ω-b in a logarithmic graph for each acceleration ramp used in our experiments and extracting their slope. Curves are shown in Figure 7 only for ω < 6000 rpm since the asymptotic trend of T versus ω toward a finite final thickness is not accounted for in Meyerhofer’s model,20 thus failing when that plateau has been reached. The linearity of such curves confirms the predicted exponential dependence of λBragg with b, and from their slope we can estimate the value of the spin parameter b. In Figure 8 we plot b versus γ. Interestingly, as γ increases, the spin parameter tends to b ) 0.5. After Birnie’s model, this curve shows that, for the larger γ employed, the crossover point is reached when the substrate is already rotating at the final speed. In those cases, we observe a clear correlation between λBragg and ω. Thus, we can establish that precise control of the lattice parameter is attained by varying exclusively ω if the thinning of each layer in the 1DPC due to

CalVo et al.

Figure 8. Spin parameter, b, versus ramp acceleration, γ. fluid flow caused by substrate rotation is still larger than the thinning due to solvent evaporation when the final rotation speed is reached. In contrast, low values of b are attained when small γ are used. For those cases we observe that λBragg depends strongly on γ (see Figure 7). In such cases, the final thickness of each layer in the 1DPC is achieved before the substrate reaches the final rotation speed, i.e., the crossover occurs during the ramp stage.

Conclusions A detailed study of the structural and optical quality of nanoparticle-based one-dimensional photonic crystals made by spin coating has been presented. We have shown that precise control over the optical response can be achieved through the variation of the spin-coating protocol. It also provides a means to reduce the number of imperfections inherent to the process of spin coating of colloidal suspensions. Based on an analysis of the spin parameter, we could relate the structural variations observed with the different relative importance of fluid flow and solvent evaporation on the thinning of each layer in the stack for the different deposition conditions employed. Acknowledgment. This research has been funded by the Spanish Ministry of Science and Education. M. E. Calvo and S. Colodrero thank CSIC for an I3P contract and scholarship, respectively. LA8030057